基于图像矩和小波能量的红外目标特征提取

Feature Extraction of Infrared Target Based on Image Moment and Wavelet Energy

特征提取是目标分类和识别的基础。图像矩反映了物体灰度相对于质心的统计分布情况。目标的红外图像信息在小波域主要集中在低频部分。文章中提出了一种提取红外目标不变特征的新方法,利用该方法提取的特征向量,对红外目标图像进行了分类识别。实验结果表明:这种方法对精确制导武器的目标识别具有较高的实用价值。

Feature extraction builds up the basis of target classification and recognition. The image moment represents the statistical distribution of gray of the image relative to the center of its mass. The infrared image information of the target mainly concentrates on the low frequency part after being transformed by wavelet. Advances a new kind of algorithm to get some invariant features of the infrared target and uses these invariant eigenvector of the method extraction to classify and recognise the infrared target image. The experimental result shows that the algorithm has valuable application in target recognition of the precision guidance weapon.